Ron Heiby (The Moderator) <unix-request@cbosgd.UUCP> (03/16/85)
Unix Technical Digest Fri, 15 Mar 85 Volume 1 : Issue 26
Today's Topics:
P-E and float?
Question on VAX Page Table Entry in 4.2BSD
Requesting info on UNIX look-alikes
Snobol for Unix?
What's wrong with this picture (makefile bug?)
----------------------------------------------------------------------
Date: 13 Mar 85 23:13:09 GMT
From: hugh@BRL-BMD.ARPA (USAFAS)
Subject: P-E and float?
Hi,
We have been given access to a Perkin-Elmer running the
Wollogong(sp?) implementation of unix and have found that the simplest
versions of programs written in both FORTRAN and C don't work.
I think that I have isolated the problem: my conclusion is that
this implementation doesn't have float in C and doesn't have
REAL in FORTRAN. When I declare my variables as double or DOUBLE
PRECISION in C or FORTRAN, respectively, the programs work; declared
as float or REAL, they don't work.
Is this standard for this implemetation? Or should I tell
the system administrator to get to work straightening out this problem?
thanks,
hugh@brl
------------------------------
Date: 13 Mar 85 00:47:33 GMT
From: danc%sri-tsca@turtlevax.UUCP
Subject: Question on VAX Page Table Entry in 4.2BSD
In machine/pte.h there are two fields in the VAX "struct pte" that
don't seem to be used for anything. They are "pg_vreadm" and "pg_swapm".
From looking at vm_mem.c and vm_page.c, it seems that pg_vreadm was
a remnant from something Berkeley never got around to doing. It
relates to "pg_fod" (Fill-on-demand) and "pg_fileno" in "struct fpte".
My guess is pg_vreadm has something to do with being able to
pagein text/data from files other than the file that was exec-ed
(maybe part of a shared-memory scheme or something)???
I have no idea what pg_swapm was intended for. Anyone know for
sure what the origin and/or use of these fields are? We are porting
the virtual-memory code to a 68020, and I don't want to eliminate them
if 4.3BSD (or whatever they plan to call it) makes use of them. Thanks!
-Dan Chernikoff
Cadlinc, Inc
Arpanet: dan@sri-tsc
Usenet: {decwrl,hplabs}!turtlevax!danc
------------------------------
Date: 14 Mar 85 00:25:50 GMT
From: wales@UCLA-LOCUS.ARPA (Rich Wales)
Subject: Requesting info on UNIX look-alikes
I would like to find out as much as possible about any UNIX look-alikes.
By "look-alike" I mean a system whose program interface looks like UNIX
(be it 4.1BSD, 4.2BSD, System III, System V, or whatever) -- but whose
internals have been rewritten from the ground up and are not based on
AT&T or Berkeley sources.
Hence, I am NOT asking for information about "ports" of UNIX code to new
machines. UNIX look-alikes which run on top of another operating system
are interesting to me, though -- as long as the look-alike was written
from scratch and is not simply a "port" of AT&T or Berkeley code.
I am primarily interested in look-alike KERNELS. If you know of a sys-
tem where the kernel has been rewritten but the AT&T/Berkeley library
routines (such as /lib/libc.a) have not been rewritten, I would still
like to know about that system.
For any such system which you know about, I would like to get at least
the following information:
(1) Name of the system, and the company or group which developed it.
(2) When was it developed? How long did it take to develop it?
(3) What hardware does it run on?
(4) What are its significant internal differences from AT&T or Berkeley
UNIX? Generalities in the spirit of the following are good enough
(I'm certainly not asking anyone to divulge trade secrets!):
(a) "The scheduler supports real-time processing."
(b) "Device drivers can be installed dynamically, a` la VMS."
(c) "Dynamically linked run-time libraries are supported."
(5) Is it object-code compatible with any AT&T or Berkeley version of
UNIX? If it is not completely compatible, what doesn't it do?
(6) Are there any significant capabilities of the look-alike which are
not in AT&T or Berkeley UNIX systems? (The answer to this question
may admittedly overlap that of question #4 above.)
(7) How does the performance of the look-alike compare with an AT&T or
Berkeley UNIX version for the same hardware (if one is available)?
(8) Is the rewritten source available, or is the look-alike sold only
in object form?
(9) How much does it cost? (If available in both source and object
forms, what is the cost for each?)
(10) Does the look-alike's vendor require that the purchaser of the
look-alike have a source-code UNIX license from AT&T or Berkeley?
If not, is an object-code license required instead?
Please reply DIRECTLY TO ME; I will post a summary to the net in a
couple of weeks.
Oh, I almost forgot --
UNIX is a trademark of Bell Labs, AT&T, or somebody like that.
VMS is a trademark of Digital Equipment Corporation.
Berkeley is a trademark of the Regents of the University of
California. :-}
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Rich Wales
University of California, Los Angeles (UCLA)
Computer Science Department
3531 Boelter Hall
Los Angeles, California 90024 // USA
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Phone: (213) 825-5683 // +1 213 825 5683
ARPANET: wales@UCLA-LOCUS.ARPA
UUCP: ...!{cepu,ihnp4,trwspp,ucbvax}!ucla-cs!wales
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
------------------------------
Date: 13 Mar 85 21:56:33 GMT
From: landwehr@nrl-css.ARPA (Carl Landwehr)
Subject: Snobol for Unix?
I am interested in obtaining a version of SNOBOL (SNOBOL4 or better) that
will run on Berkeley UNIX on a VAX (or a SUN). I have been out of touch
with SNOBOL for quite a while and would appreciate any pointers people
can provide. Please respond to:
landwehr at NRL-CSS
Thanks much.
--Carl Landwehr
------------------------------
Date: 13 Mar 85 18:45:02 GMT
From: eric@osiris.UUCP (Eric Bergan)
Subject: What's wrong with this picture (makefile bug?)
I seem to have run into a strange "make" problem involving user
defined suffixes. I started with a makefile:
.SUFFIXES: .qc
.qc.c:
cat $< >mtest.c
mtest: mtest.o
cc -o mtest mtest.o
and a file "mtest.qc". If I say "make mtest" - it comes back "don't know
how to make mtest.o". However if I do "make mtest.c" and then "make mtest"
it works fine.
If I change the makefile to:
.SUFFIXES: .qc
.qc.c:
cat $< >mtest.c
mtest: mtest.o
cc -o mtest mtest.o
mtest.o: mtest.c
and the only file is "mtest.qc", and I do a "make mtest", it works fine.
This happens under both 4.2 and System V. Has anyone gotten user defined
suffixes to work correctly?
--
eric
...!seismo!umcp-cs!aplvax!osiris!eric
------------------------------
End of Unix Technical Digest
******************************
--
Ronald W. Heiby / ihnp4!{wnuxa!heiby|wnuxb!netnews}
AT&T Information Systems, Inc.
Lisle, IL (CU-D21)